ACADEMY

MINING EDITION

MODULE 3

LESSON 3

Advanced Miner Configuration (Qubic.li Client & Alienminer)

By now, you likely have the basics of mining with the Qubic.li client (the standard miner). In this lesson, we’ll explore advanced configuration options that allow you to fine-tune how your miner runs. We’ll also introduce Alienminer, an unofficial high-performance mining software favored by some advanced users, and discuss benchmarks, custom flags, and logging.

Qubic.li Miner (QLI) Advanced Settings: The Qubic.li client is designed to be user-friendly, but it offers plenty of configurability under the hood . You can run it via command-line flags, environment variables, or a config file (appsettings.json). If you made a Qubic.li pool account, the easiest route is to use the Web Client Setup Configurator on the pool site, which generates a ready-made config string or token for you . This configurator lets you pick CPU vs GPU mining, number of threads, and whether to enable Monero idle mining, then provides an Access Token or config that you plug into the miner . For example, you might get a one-line command like:

./qli-Client --ClientSettings:QubicAddress=<Your_Qubic_Wallet>
--ClientSettings:Trainer:CpuThreads

This would launch the miner pointing at your Qubic address with N CPU threads . Alternatively, you can save these settings in appsettings.json in the same folder as the miner. An example config for a 10-thread CPU miner might look like:

{
  "ClientSettings": {
    "accessToken": "YOUR_ACCESS_TOKEN_OR_ADDRESS",
    "alias": "MyRig1",
    "trainer": { "cpu": true, "cpuThreads": 10 }
  }
}

This file approach is convenient for persistent rigs – just double-click the client and it will read the JSON. The Qubic.li client also supports Docker and even a systemd service on Linux for running in the background (ideal for headless servers). Advanced users can set up the miner to auto-start at boot using these methods, ensuring you’re always online.

Config Flags to Know

Some useful QLI config flags include CpuThreads (how many CPU threads to use) , gpu and gpuVersion (to enable GPU mining and specify CUDA vs AMD) , Alias (a custom name for your worker, shows on the pool dashboard) , and pps (true/false for Pay-Per-Share vs solo mode in the pool) . Pay-Per-Share (PPS+) means you get steady income based on work contributed, smoothing out luck . Solo means you only get reward if your miner actually finds a valid solution, which is higher variance but potentially higher payout if you’re powerful – Qubic.li even allows a mix (some threads PPS, some solo) for flexibility . Most beginners stick to PPS, but advanced miners might go solo if they have a lot of hashpower and want to avoid pool fees. You can toggle this in the config JSON or via the web configurator easily.

Idle Mining (XMR) Settings

By default, the Qubic.li client’s Monero mining module is enabled and will kick in during idle periods . Advanced users can choose to disable this if, for example, you prefer your own method of using idle time or want to reduce power usage when no Qubic tasks are available. In the config, the xmrSettings block has a flag disable (set it to true to turn off XMR mining) . You can also point the XMR miner to different pool endpoints if needed – Qubic.li provides regional XMR stratum addresses for optimal latency (EU, US, HK, SG, etc.) , but typically you won’t need to touch this. Note that currently GPU mining for XMR is not supported in the client (it’s CPU only for the idle mining) . The XMR module uses a variant of XMRig under the hood, and you could pass custom parameters to it (e.g., limit it to 1 thread) via the customParameters field . However, it’s generally recommended not to modify the XMR settings unless you really know what you’re doing – the defaults are optimized to convert idle time into extra QUBIC smoothly .

Alienminer – the Performance-Tuned Miner

For those seeking maximum performance (and who are comfortable with more complexity), the Qubic community has developed Alienminer, an unofficial mining software . Alienminer is focused on squeezing out extra hashes beyond what the standard Qubic.li client might achieve . It may include low-level optimizations, custom kernels, or fine-grained control over the training process. Some miners report higher iterations per second with Alienminer, especially on certain hardware. However, be cautious: Alienminer is not maintained by the official Qubic team, and it requires manual updates when Qubic epochs or algorithms change (the Qubic.li client auto-downloads new neural network “trainers” each epoch; with Alienminer you might need to update these yourself). It’s also primarily distributed via community channels (such as Discord), so you should only download it from trusted sources. The Qubic Docs note that Alienminer “may offer features and optimizations not found in the Qubic.li client” , but it’s recommended only for advanced users who can handle potential issues. If you try Alienminer, keep your Qubic.li client handy as a fallback.

Benchmarking & Logging

How do you know if your tweaks are helping? Both Qubic.li and Alienminer typically display performance info in the console (e.g., iterations per second, shares found, etc.). If you have a pool account, the pool dashboard also shows your current hash rate (though in Qubic’s case it might show as “training performance” rather than a hash hash rate) and your contribution relative to others. A good practice is to change one setting at a time and let the miner run for a while (several minutes at least) to gauge the steady-state performance. For instance, try CpuThreads=16 vs CpuThreads=32 and compare the output. The logging in Qubic.li client is fairly minimal by default (mostly status updates and any warnings/errors). If something goes wrong, the client will drop error logs in the log/ folder with timestamps . Advanced users can increase verbosity by running the client in a console to see all messages, or by enabling debug flags if available (check Qubic’s GitHub for any debug build instructions). The legacy Computor software had function keys to toggle detailed logging (F7, F8, F9) , but in the current pool miner those aren’t used. Instead, rely on the log files for troubleshooting (we’ll cover common issues in Lesson 5). If you’re running Alienminer, be prepared for a more bare-bones output – you might need to manually update config files or even recompile for new epochs. Always document your changes and results so you can roll back if a “tweak” makes things worse.

PRO TIP

Consistently monitor your rig. Use tools like htop or Task Manager to ensure no other background processes are stealing CPU cycles that could be mining. On GPUs, avoid running UI/monitors on the same card that’s mining if possible – for example, if your CPU has an integrated GPU, use that for display output and dedicate your discrete GPU entirely to Qubic . This prevents Windows or X11 from causing stutters or lost mining time on the main GPU. Little tweaks like these ensure Qubic’s useful work gets almost all of your hardware’s attention.

PRO TIP ADVANCED

If you have a high-core-count CPU with AVX-512 support, you might find the chip downclocking under AVX-512 load. In some cases, forcing the miner to use the AVX2 trainer instead can yield higher overall throughput (due to higher sustained clock speeds). You can do this by setting "cpuVersion": "AVX2" in your config overwrites . Conversely, if your CPU supports a special mode like Intel’s Skylake-X, you could enforce "cpuVersion": "SKYLAKE" . These options are there if the auto-detection ever picks a suboptimal trainer for your CPU. Use them carefully and only if you notice an issue.

PRO TIP

When fine-tuning, track your weekly performance score on the pool. Because Qubic rewards are weekly, a change that gives you +5% iterations per second will translate to ~5% more QUBIC by week’s end, which is significant! Use the pool’s charts or your own logs to measure improvements. Also, don’t neglect stability: a super aggressive setting that crashes your miner after an hour will lose you more time than it gains. It’s better to have a stable, 24/7 running configuration at 95% of max performance than a flaky one at 100%.

PRO TIP ADVANCED

If you have a high-core-count CPU with AVX-512 support, you might find the chip downclocking under AVX-512 load. In some cases, forcing the miner to use the AVX2 trainer instead can yield higher overall throughput (due to higher sustained clock speeds). You can do this by setting "cpuVersion": "AVX2" in your config overwrites . Conversely, if your CPU supports a special mode like Intel’s Skylake-X, you could enforce "cpuVersion": "SKYLAKE" . These options are there if the auto-detection ever picks a suboptimal trainer for your CPU. Use them carefully and only if you notice an issue.

© 2025 Qubic.

Qubic is a decentralized, open-source network for experimental technology. Nothing on this site should be construed as investment, legal, or financial advice. Qubic does not offer securities, and participation in the network may involve risks. Users are responsible for complying with local regulations. Please consult legal and financial professionals before engaging with the platform.

© 2025 Qubic.

Qubic is a decentralized, open-source network for experimental technology. Nothing on this site should be construed as investment, legal, or financial advice. Qubic does not offer securities, and participation in the network may involve risks. Users are responsible for complying with local regulations. Please consult legal and financial professionals before engaging with the platform.